html2rss 0.15.0 → 0.17.0

data/lib/html2rss/auto_source/scraper/schema/thing.rb

@@ -0,0 +1,104 @@
+# frozen_string_literal: true
+
+require 'date'
+
+module Html2rss
+  class AutoSource
+    module Scraper
+      class Schema
+        ##
+        # A Thing is kind of the 'base class' for Schema.org schema_objects.
+        #
+        # @see https://schema.org/Thing
+        class Thing
+          SUPPORTED_TYPES = %w[
+            AdvertiserContentArticle
+            AnalysisNewsArticle
+            APIReference
+            Article
+            AskPublicNewsArticle
+            BackgroundNewsArticle
+            BlogPosting
+            DiscussionForumPosting
+            LiveBlogPosting
+            NewsArticle
+            OpinionNewsArticle
+            Report
+            ReportageNewsArticle
+            ReviewNewsArticle
+            SatiricalArticle
+            ScholarlyArticle
+            SocialMediaPosting
+            TechArticle
+          ].to_set.freeze
+
+          DEFAULT_ATTRIBUTES = %i[id title description url image published_at].freeze
+
+          def initialize(schema_object, url:)
+            @schema_object = schema_object
+            @url = url
+          end
+
+          # @return [Hash] the scraped article hash with DEFAULT_ATTRIBUTES
+          def call
+            DEFAULT_ATTRIBUTES.to_h do |attribute|
+              [attribute, public_send(attribute)]
+            end
+          end
+
+          def id
+            return @id if defined?(@id)
+
+            id = (schema_object[:@id] || url&.path).to_s
+
+            return if id.empty?
+
+            @id = id
+          end
+
+          def title = schema_object[:title]
+
+          def description
+            schema_object.values_at(:description, :schema_object_body, :abstract)
+                         .max_by { |string| string.to_s.size }
+          end
+
+          # @return [Addressable::URI, nil] the URL of the schema object
+          def url
+            url = schema_object[:url]
+            if url.to_s.empty?
+              Log.debug("Schema#Thing.url: no url in schema_object: #{schema_object.inspect}")
+              return
+            end
+
+            Utils.build_absolute_url_from_relative(url, @url)
+          end
+
+          def image
+            if (image_url = image_urls.first)
+              Utils.build_absolute_url_from_relative(image_url, @url)
+            end
+          end
+
+          def published_at = schema_object[:datePublished]
+
+          private
+
+          attr_reader :schema_object
+
+          def image_urls
+            schema_object.values_at(:image, :thumbnailUrl).filter_map do |object|
+              next unless object
+
+              if object.is_a?(String)
+                object
+              elsif object.is_a?(Hash) && object[:@type] == 'ImageObject'
+                object[:url] || object[:contentUrl]
+              end
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/lib/html2rss/auto_source/scraper/schema.rb

@@ -2,58 +2,38 @@
 
 require 'json'
 require 'nokogiri'
-require 'set'
 
 module Html2rss
   class AutoSource
     module Scraper
       ##
-      # Scraps articles from Schema.org objects, by looking for the objects in:
+      # Scrapes articles from Schema.org objects, by looking for the objects in:
 
-      #  1. <script type="application/ld+json"> "schema" tag.
-      #  2. tbd
+      # <script type="application/ld+json"> "schema" tags.
       #
       # See:
-      # 1. https://schema.org/NewsArticle
+      # 1. https://schema.org/docs/full.html
       # 2. https://developers.google.com/search/docs/appearance/structured-data/article#microdata
       class Schema
         include Enumerable
 
         TAG_SELECTOR = 'script[type="application/ld+json"]'
-        SCHEMA_OBJECT_TYPES = %w[
-          AdvertiserContentArticle
-          AnalysisNewsArticle
-          APIReference
-          Article
-          AskPublicNewsArticle
-          BackgroundNewsArticle
-          BlogPosting
-          DiscussionForumPosting
-          LiveBlogPosting
-          NewsArticle
-          OpinionNewsArticle
-          Report
-          ReportageNewsArticle
-          ReviewNewsArticle
-          SatiricalArticle
-          ScholarlyArticle
-          SocialMediaPosting
-          TechArticle
-        ].to_set.freeze
 
         class << self
           def articles?(parsed_body)
             parsed_body.css(TAG_SELECTOR).any? do |script|
-              SCHEMA_OBJECT_TYPES.any? { |type| script.text.match?(/"@type"\s*:\s*"#{Regexp.escape(type)}"/) }
+              (Thing::SUPPORTED_TYPES | ItemList::SUPPORTED_TYPES).any? do |type|
+                script.text.match?(/"@type"\s*:\s*"#{Regexp.escape(type)}"/)
+              end
             end
           end
 
           ##
           # Returns a flat array
           # of all supported schema objects
-          # by recursively traversing the `from` object.
+          # by recursively traversing the given `object`.
           #
-          # @param object [Hash, Array]
+          # @param object [Hash, Array, Nokogiri::XML::Element]
           # @return [Array<Hash>] the schema_objects, or an empty array
           # :reek:DuplicateMethodCall
           def from(object)
@@ -74,12 +54,16 @@ module Html2rss
           end
 
           ##
-          # @return [Scraper::Schema::Base, Scraper::Schema::NewsArticle, nil]
+          # @return [Scraper::Schema::Thing, Scraper::Schema::ItemList, nil] a class responding to `#call`
           def scraper_for_schema_object(schema_object)
-            if SCHEMA_OBJECT_TYPES.member?(schema_object[:@type])
-              Base
+            type = schema_object[:@type]
+
+            if Thing::SUPPORTED_TYPES.member?(type)
+              Thing
+            elsif ItemList::SUPPORTED_TYPES.member?(type)
+              ItemList
             else
-              Log.warn("Schema#scraper_for_schema_object: Unsupported schema object @type: #{schema_object[:@type]}")
+              Log.warn("Schema#scraper_for_schema_object: Unsupported schema object @type: #{type}")
               nil
             end
           end
@@ -107,9 +91,13 @@ module Html2rss
 
           schema_objects.filter_map do |schema_object|
             next unless (klass = self.class.scraper_for_schema_object(schema_object))
-            next unless (article_hash = klass.new(schema_object, url:).call)
+            next unless (results = klass.new(schema_object, url:).call)
 
-            yield article_hash
+            if results.is_a?(Array)
+              results.each { |result| yield(result) } # rubocop:disable Style/ExplicitBlockArgument
+            else
+              yield(results)
+            end
           end
         end
 

data/lib/html2rss/auto_source/scraper/semantic_html/extractor.rb

@@ -1,7 +1,5 @@
 # frozen_string_literal: true
 
-require 'set'
-
 module Html2rss
   class AutoSource
     module Scraper
@@ -33,6 +31,8 @@ module Html2rss
           end
 
           def initialize(article_tag, url:)
+            raise ArgumentError, 'article_tag is required' unless article_tag
+
             @article_tag = article_tag
             @url = url
           end
@@ -57,20 +57,6 @@ module Html2rss
 
           attr_reader :article_tag, :url, :heading, :extract_url
 
-          def visible_text_from_tag(tag, separator: ' ') = self.class.visible_text_from_tag(tag, separator:)
-
-          # @see https://developer.mozilla.org/en-US/docs/Web/API/HTMLTimeElement/dateTime
-          def extract_published_at
-            times = article_tag.css('time[datetime]')
-                               .filter_map do |tag|
-              DateTime.parse(tag['datetime'])
-            rescue ArgumentError, TypeError
-              nil
-            end
-
-            times.min
-          end
-
           ##
           # Find the heading of the article.
           # @return [Nokogiri::XML::Node, nil]
@@ -80,18 +66,36 @@ module Html2rss
             return if heading_tags.empty?
 
             smallest_heading = heading_tags.keys.min
-            heading_tags[smallest_heading]&.max_by { |tag| visible_text_from_tag(tag)&.size }
+            heading_tags[smallest_heading]&.max_by { |tag| visible_text_from_tag(tag)&.size.to_i }
+          end
+
+          def visible_text_from_tag(tag, separator: ' ') = self.class.visible_text_from_tag(tag, separator:)
+
+          def closest_anchor
+            SemanticHtml.find_closest_selector(heading || article_tag,
+                                               selector: 'a[href]:not([href=""])')
+          end
+
+          def find_url
+            href = closest_anchor&.[]('href')
+
+            return if (parts = href.to_s.split('#')).empty?
+
+            Utils.build_absolute_url_from_relative(parts.first.strip, url)
           end
 
           def extract_title
-            @extract_title ||= if heading && (heading.children.empty? || heading.text)
-                                 visible_text_from_tag(heading)
-                               else
-                                 visible_text_from_tag(
-                                   article_tag.css(HEADING_TAGS.join(','))
-                                              .max_by { |tag| tag.text.size }
-                                 )
-                               end
+            if heading && (heading.children.empty? || heading.text)
+              visible_text_from_tag(heading)
+            else
+              visible_text_from_tag(article_tag.css(HEADING_TAGS.join(','))
+                                               .max_by { |tag| tag.text.size })
+
+            end
+          end
+
+          def extract_image
+            Image.call(article_tag, url:)
           end
 
           def extract_description
@@ -101,26 +105,10 @@ module Html2rss
             description = visible_text_from_tag(article_tag)
             return nil unless description
 
-            title_text = extract_title
-            description.gsub!(title_text, '') if title_text
             description.strip!
             description.empty? ? nil : description
           end
 
-          def closest_anchor
-            SemanticHtml.find_closest_selector(heading || article_tag,
-                                               selector: 'a[href]:not([href=""])')
-          end
-
-          def find_url
-            href = closest_anchor&.[]('href')&.split('#')&.first&.strip
-            Utils.build_absolute_url_from_relative(href, url) unless href.to_s.empty?
-          end
-
-          def extract_image
-            Image.call(article_tag, url:)
-          end
-
           def generate_id
             [
               article_tag['id'],
@@ -129,6 +117,18 @@ module Html2rss
               extract_url&.query
             ].compact.reject(&:empty?).first
           end
+
+          # @see https://developer.mozilla.org/en-US/docs/Web/API/HTMLTimeElement/dateTime
+          def extract_published_at
+            times = article_tag.css('time[datetime]')
+                               .filter_map do |tag|
+              DateTime.parse(tag['datetime'])
+            rescue ArgumentError, TypeError
+              nil
+            end
+
+            times.min
+          end
         end
       end
     end

data/lib/html2rss/auto_source/scraper/semantic_html/image.rb

@@ -28,14 +28,14 @@ module Html2rss
           # @see <https://developer.mozilla.org/en-US/docs/Web/HTML/Element/picture>
           def self.from_source(article_tag) # rubocop:disable Metrics/AbcSize
             hash = article_tag.css('img[srcset], picture > source[srcset]')
-                              .flat_map { |source| source['srcset'].to_s.split(',') }
-                              .filter_map do |line|
-              width, url = line.split.reverse
-              next if url.nil? || url.start_with?('data:')
+                              .flat_map do |source|
+              source['srcset'].to_s.scan(/(\S+)\s+(\d+w|\d+h)/).map do |url, width|
+                next if url.nil? || url.start_with?('data:')
 
-              width_value = width.to_i.zero? ? 0 : width.scan(/\d+/).first.to_i
+                width_value = width.to_i.zero? ? 0 : width.scan(/\d+/).first.to_i
 
-              [width_value, url.strip]
+                [width_value, url.strip]
+              end
             end.to_h
 
             hash[hash.keys.max]

data/lib/html2rss/auto_source/scraper/semantic_html.rb

@@ -106,9 +106,10 @@ module Html2rss
           SemanticHtml.anchor_tag_selector_pairs.each do |tag_name, selector|
             parsed_body.css(selector).each do |selected_tag|
               article_tag = SemanticHtml.find_tag_in_ancestors(selected_tag, tag_name)
-              article_hash = Extractor.new(article_tag, url: @url).call
 
-              yield article_hash if article_hash
+              if article_tag && (article_hash = Extractor.new(article_tag, url: @url).call)
+                yield article_hash
+              end
             end
           end
         end

data/lib/html2rss/auto_source.rb

@@ -11,20 +11,13 @@ module Html2rss
   # It uses a set of ArticleExtractors to extract articles, utilizing popular ways of
   # marking articles, e.g. schema, microdata, open graph, etc.
   class AutoSource
-    class UnsupportedUrlScheme < Html2rss::Error; end
     class NoArticlesFound < Html2rss::Error; end
 
-    SUPPORTED_URL_SCHEMES = %w[http https].to_set.freeze
-
     ##
     # @param url [Addressable::URI] The URL to extract articles from.
     # @param body [String] The body of the response.
     # @param headers [Hash] The headers of the response.
     def initialize(url, body:, headers: {})
-      raise ArgumentError, 'URL must be a Addressable::URI' unless url.is_a?(Addressable::URI)
-      raise ArgumentError, 'URL must be absolute' unless url.absolute?
-      raise UnsupportedUrlScheme, "#{url.scheme} not supported" unless SUPPORTED_URL_SCHEMES.include?(url.scheme)
-
       @url = url
       @body = body
       @headers = headers

data/lib/html2rss/cli.rb

@@ -2,7 +2,6 @@
 
 require_relative '../html2rss'
 require 'thor'
-require 'addressable'
 
 ##
 # The Html2rss namespace / command line interface.
@@ -26,14 +25,22 @@ module Html2rss
     def feed(yaml_file, *options)
       raise "File '#{yaml_file}' does not exist" unless File.exist?(yaml_file)
 
-      feed_name = options.shift
+      feed_name = options.shift unless options.first&.include?('=')
       params = options.to_h { |opt| opt.split('=', 2) }
+
       puts Html2rss.feed_from_yaml_config(yaml_file, feed_name, params:)
     end
 
-    desc 'auto URL', 'automatically sources an RSS feed from the URL'
+    desc 'auto URL', 'Automatically sources an RSS feed from the URL'
+    method_option :strategy,
+                  type: :string,
+                  desc: 'The strategy to request the URL',
+                  enum: RequestService.strategy_names,
+                  default: RequestService.default_strategy_name
     def auto(url)
-      puts Html2rss.auto_source(url)
+      strategy = options.fetch(:strategy) { RequestService.default_strategy_name }.to_sym
+
+      puts Html2rss.auto_source(url, strategy:)
     end
   end
 end

data/lib/html2rss/config/channel.rb

@@ -55,7 +55,7 @@ module Html2rss
       ##
       # @return [String]
       def title
-        config.fetch(:title) { Utils.titleized_url(url) }
+        config.fetch(:title) { Utils.titleized_channel_url(url) }
       end
 
       ##
@@ -88,6 +88,12 @@ module Html2rss
         config.fetch(:json, false)
       end
 
+      ##
+      # @return [Symbol]
+      def strategy
+        config.fetch(:strategy) { RequestService.default_strategy_name }.to_sym
+      end
+
       private
 
       # @return [Hash<Symbol, Object>]

data/lib/html2rss/config/selectors.rb

@@ -8,7 +8,8 @@ module Html2rss
       ITEMS_SELECTOR_NAME = :items
 
       # Struct to represent a selector with associated attributes for extraction and processing.
-      Selector = Struct.new(:selector, :attribute, :extractor, :post_process, :order, :static, keyword_init: true)
+      Selector = Struct.new(:selector, :attribute, :extractor, :post_process, :order, :static, :content_type,
+                            keyword_init: true)
 
       # raised when an invalid selector name is used
       class InvalidSelectorName < Html2rss::Error; end

data/lib/html2rss/config.rb

@@ -27,6 +27,7 @@ module Html2rss
     def_delegator :@channel, :url, :link
     def_delegator :@channel, :time_zone
     def_delegator :@channel, :json?
+    def_delegator :@channel, :strategy
 
     def_delegator :@selectors, :item_selector_names
     def_delegator :@selectors, :selector?

data/lib/html2rss/item.rb

@@ -23,7 +23,9 @@ module Html2rss
     # @param config [Html2rss::Config] Configuration object.
     # @return [Array<Html2rss::Item>] list of items fetched.
     def self.from_url(url, config)
-      body = Utils.request_url(url, headers: config.headers).body
+      ctx = RequestService::Context.new(url:, headers: config.headers)
+
+      body = RequestService.execute(ctx, strategy: config.strategy).body
       body = ObjectToXmlConverter.new(JSON.parse(body)).call if config.json?
 
       Nokogiri.HTML(body)
@@ -136,8 +138,11 @@ module Html2rss
 
       raise 'An item.enclosure requires an absolute URL' unless url&.absolute?
 
+      type = config.selector_attributes_with_channel(:enclosure)[:content_type] ||
+             Html2rss::Utils.guess_content_type_from_url(url)
+
       Enclosure.new(
-        type: Html2rss::Utils.guess_content_type_from_url(url),
+        type:,
         bits_length: 0,
         url: url.to_s
       )

data/lib/html2rss/request_service/browserless_strategy.rb

@@ -0,0 +1,53 @@
+# frozen_string_literal: true
+
+require 'puppeteer'
+
+module Html2rss
+  class RequestService
+    ##
+    # Browserless.io strategy to request websites.
+    #
+    # Provide the WebSocket URL and your API token via environment variables:
+    # - BROWSERLESS_IO_WEBSOCKET_URL
+    # - BROWSERLESS_IO_API_TOKEN
+    #
+    # To use this strategy, you need to have a Browserless.io account or run a
+    # local Browserless.io instance.
+    #
+    # @see https://www.browserless.io/
+    #
+    # To run a local Browserless.io instance, you can use the following Docker command:
+    #
+    # ```sh
+    # docker run \
+    #   --rm \
+    #   -p 3000:3000 \
+    #   -e "CONCURRENT=10" \
+    #   -e "TOKEN=6R0W53R135510" \
+    #   ghcr.io/browserless/chromium
+    # ```
+    #
+    # When running locally, you can skip setting the environment variables, as above commands
+    # are aligned with the default values.
+    # @see https://github.com/browserless/browserless/pkgs/container/chromium
+    class BrowserlessStrategy < Strategy
+      # return [Response]
+      def execute
+        Puppeteer.connect(browser_ws_endpoint:) do |browser|
+          PuppetCommander.new(ctx, browser).call
+        ensure
+          browser.disconnect
+        end
+      end
+
+      def browser_ws_endpoint
+        @browser_ws_endpoint ||= begin
+          api_token = ENV.fetch('BROWSERLESS_IO_API_TOKEN', '6R0W53R135510')
+          ws_url = ENV.fetch('BROWSERLESS_IO_WEBSOCKET_URL', 'ws://127.0.0.1:3000')
+
+          "#{ws_url}?token=#{api_token}"
+        end
+      end
+    end
+  end
+end

data/lib/html2rss/request_service/context.rb

@@ -0,0 +1,46 @@
+# frozen_string_literal: true
+
+require 'addressable/uri'
+
+module Html2rss
+  class RequestService
+    ##
+    # Holds information needed to send requests to websites.
+    # To be passed down to the RequestService's strategies.
+    class Context
+      SUPPORTED_URL_SCHEMES = %w[http https].to_set.freeze
+
+      ##
+      # @param url [String, Addressable::URI] the URL to request
+      # @param headers [Hash] HTTP request headers
+      def initialize(url:, headers: {})
+        @url = Addressable::URI.parse(url)
+        assert_valid_url!
+
+        @headers = headers
+      end
+
+      # @return [Addressable::URI] the parsed URL
+      attr_reader :url
+
+      # @return [Hash] the HTTP request headers
+      attr_reader :headers
+
+      private
+
+      ##
+      # Validates the URL.
+      # @raise [InvalidUrl] if the URL is not valid
+      # @raise [UnsupportedUrlScheme] if the URL scheme is not supported
+      def assert_valid_url!
+        raise InvalidUrl, 'URL must be absolute' unless url.absolute?
+        raise InvalidUrl, 'URL must not contain an @ character' if url.to_s.include?('@')
+
+        return if SUPPORTED_URL_SCHEMES.include?(url.scheme)
+
+        raise UnsupportedUrlScheme,
+              "URL scheme '#{url.scheme}' is not supported"
+      end
+    end
+  end
+end

data/lib/html2rss/request_service/faraday_strategy.rb

@@ -0,0 +1,24 @@
+# frozen_string_literal: true
+
+require 'faraday'
+require 'faraday/follow_redirects'
+
+module Html2rss
+  class RequestService
+    ##
+    # Strategy to use Faraday for the request.
+    # @see https://rubygems.org/gems/faraday
+    class FaradayStrategy < Strategy
+      # return [Response]
+      def execute
+        request = Faraday.new(url: ctx.url, headers: ctx.headers) do |faraday|
+          faraday.use Faraday::FollowRedirects::Middleware
+          faraday.adapter Faraday.default_adapter
+        end
+        response = request.get
+
+        Response.new(body: response.body, headers: response.headers)
+      end
+    end
+  end
+end

data/lib/html2rss/request_service/puppet_commander.rb

@@ -0,0 +1,61 @@
+# frozen_string_literal: true
+
+module Html2rss
+  class RequestService
+    ##
+    # Commands the Puppeteer Browser to the website and builds the Response.
+    class PuppetCommander
+      # @param ctx [Context]
+      # @param browser [Puppeteer::Browser]
+      # @param skip_request_resources [Set<String>] the resource types not to request
+      # @param referer [String] the referer to use for the request
+      def initialize(ctx,
+                     browser,
+                     skip_request_resources: %w[stylesheet image media font].to_set,
+                     referer: [ctx.url.scheme, ctx.url.host].join('://'))
+        @ctx = ctx
+        @browser = browser
+        @skip_request_resources = skip_request_resources
+        @referer = referer
+      end
+
+      # @return [Response]
+      def call
+        page = new_page
+
+        response = navigate_to_destination(page, ctx.url)
+
+        Response.new(body: body(page), headers: response.headers)
+      ensure
+        page&.close
+      end
+
+      ##
+      # @return [Puppeteer::Page]
+      # @see https://yusukeiwaki.github.io/puppeteer-ruby-docs/Puppeteer/Page.html
+      def new_page
+        page = browser.new_page
+        page.extra_http_headers = ctx.headers
+
+        return page if skip_request_resources.empty?
+
+        page.request_interception = true
+        page.on('request') do |request|
+          skip_request_resources.member?(request.resource_type) ? request.abort : request.continue
+        end
+
+        page
+      end
+
+      def navigate_to_destination(page, url)
+        page.goto(url, wait_until: 'networkidle0', referer:)
+      end
+
+      def body(page) = page.content
+
+      private
+
+      attr_reader :ctx, :browser, :skip_request_resources, :referer
+    end
+  end
+end